Re: new or old devfs naming scheme (Was: Re: devfs howto)

From: Wakko Warner (wakko@animx.eu.org)
Date: Mon Feb 21 2000 - 07:25:11 EST


> >
> > Hello Richard, I have to agree with Sergey on the point
> > of liking the old names over the new . But I guess I can
> > get devfsd to create them for me can't I ? Tia, JimL
>
> Yes, you can. But they won't appear until devfsd is fired up. So you can't
> mount the root fs with old name until the devfsd which happen to be on the
> fs which you wanna mount is running... One can squeeze statically linked
> devfsd into initrd, but with modular filesystems, lvm etc. it'll be rather
> big one...
>
> Richard, is it possible and can be agreed upon with Linus to make a choice
> of new names vs old ones a kernel boot option? I'd be helluva pleased to have
> such an option. Isn't is ugly to have such a horror?
>
> === Cut ===
> [root@nomad /root]# df
> Filesystem 1k-blocks Used Available Use% Mounted on
> /dev/scsi/host0/bus0/target0/lun0/part1
> 16973507 13840592 2248940 86% /
> /dev/ide/host0/bus0/target0/lun0/part1
> 8583124 6214693 1923351 76%
> /home/ftp/ftp.ksi-linux.com
> === Cut ===

I have a question. I know this would be confusing for a user to understand
at first. Is there a way to combine scsi controllers/channels and ide's
into a single controller number? like c0 is scsi0:0 c1 is scsi0:1 and c2 is
ide0 ? As I said, it would be confusing, I'm just wondering how hard it
would be to add functionality like that into devfs. Personally, I like the
idea of this and like the way solaris handles disk names.

I remember the war over this some months ago, I do not wish to start it back
up.

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Wed Feb 23 2000 - 21:00:27 EST